Blog by Sandy Harrison
I was delighted to be spearheading a workshop held at University of Reading in July that brought together observationalists and modellers working on palaeoclimates, model development, and assessment of future climate changes to address this question.
Climate models are mathematical representations of the interactions between the atmosphere, oceans, land surface, ice – and the sun. The models used for future climate projections were developed and calibrated using climate observations from the past 40 years. They are also the only tools available to project the human impact on the environment changes over the 21st century. These models perform well in terms of global features (e.g. magnitude of global warming), but model performance at a regional scale is poor.
By researching past climates (paleoclimatology), studying the fossils of plants and animals preserved in lake or bog sediments, we can determine how ecosystems have changed over time. This has practical importance to our lives today and in the future.
Records of the past tell us what the climate system can do. For example, if there have been shifts from grasslands to forests in Northern Africa in response to major shifts in rainfall, then similar shifts can happen again. Understanding the natural climate variability, those that come about without external factors such as human impact provide a baseline from which to assess the effects of these forced changes.
Currently only very limited use has been made of the power of the palaeorecord to improve modelling capacity. In part, this is because there is a disconnect between the model-development and palaeoclimate communities. It also reflects the limited focus on evaluating earth-system components (as opposed to surface climate) by the palaeoclimate community.
An important question we looked at in-depth at this milestone Past2Future workshop is whether the relationships developed from modern observations will be reliable when we move beyond our current range of climate models. The past holds the key here and we will be looking at the potential for palaeodata to be better integrated in model development, tuning and evaluation. Most significantly, I hope this workshop will be the catalyst for the development of a new framework to improve Earth System Models.
Sandy Harrison is Professor in Global Palaeoclimates and Biogeochemical Cycles in the School of Geography and Environmental Science, University of Reading.